Real-time brachytherapy for prostate cancer implant analysis

Transcription

1 Rep Pract Oncol Radiother, 2008; 13(1): 9-14 Original Paper Received: Accepted: Published: Authors Contribution: A Study Design B Data Collection C Statistical Analysis D Data Interpretation E Manuscript Preparation F Literature Search G Funds Collection Real-time brachytherapy for prostate cancer implant analysis Marta Szlag 1 ABCDEF, Krzysztof Ślosarek 1 ACD, Agata Rembielak 2 ACD, Brygida Białas 2 AD, Marek Fijałkowski 2 B, Joanna Bystrzycka 1 B 1 Department of Treatment Planning, Maria Skłodowska-Curie Memorial Centre of Oncology Institute Gliwice Branch, Gliwice, Poland 2 Department of Brachytherapy, Maria Skłodowska-Curie Memorial Centre of Oncology Institute Gliwice Branch, Gliwice, Poland Summary Background Aim Materials/Methods Results Conclusions Key words In HDR brachytherapy precision of catheter implantation is crucial for conformal treatment planning as a starting point for better optimization process. The aim was to investigate differences between virtual and real needle position and the effect of needle displacement on dosimetric parameters as a function of prostate volume for better evaluation of real implant with respect to final dose distribution. Thirty treatment plans calculated by Nucletron SWIFT were randomly selected. Dosimetric data including V100 for prostate gland and D10 for urethra were analyzed as a function of prostate volume and needle displacement. Needle displacement was determined by measuring the distance between virtual and real positions of respective needles in three sectional images: at the base, apex of the prostate gland and reference image. Dosimetric parameters were determined for consecutive computer plans: virtual (before implantation), live (after implantation and renewed optimization). For the purpose of this study a new parameter, VD (Volume-Dose), was created. VD indicates the quality of real dose distribution with respect to virtual treatment plan. In order to realize the assumption of virtual plan (VD<1) for a given prostate volume, mean values of needle displacement r should not exceed r acceptable according to the formula: r acceptable (Vp) μ Vp 2. For larger glands (above 30cc) final dose distribution is less dependent on needle displacement than smaller ones. The experiment determined maximum values of needle displacement for a given Vp parameter, allowing one to take advantage of optimization algorithms and to improve the final dose distribution. real-time brachytherapy interstitial implant prostate volume 9

2 Original Paper Rep Pract Oncol Radiother, 2008; 13(1): 9-14 Full-text PDF: Word count: 1595 Tables: 1 Figures: 6 References: 10 Author s address: Marta Szlag, Centrum Onkologii Instytut im. Marii Skłodowskiej-Curie, Wybrzeże Armii Krajowej 15 Str., Gliwice Poland, martaszlag@o2.pl BACKGROUND High dose rate brachytherapy has become an incredibly popular method for early-stage prostate cancer [1 3]. In HDR remote afterloading technique the source is positioned inside the catheters in predefined dwell positions. Moreover, the dwell times (time of source stopovers) inside the catheter are individually selected (optimized) for better targeting of the volume of interest as well as sparing normal tissue and organs at risks [4 6]. Additionally, increasingly improved imaging techniques facilitate accurate implantation. Real-time HDR brachytherapy based on transrectal ultrasound guidance provides continuous visualization of the prostate during the treatment planning process [1]. However, in brachytherapy, precision of catheter implantation according to pre-planned needle position and patient s anatomy is crucial for conformal treatment planning and in HDR stepping source technique is a starting point for better optimization process. There is a lack of reports in the literature on the subject of needle displacement as the distance between preplanned catheter position and its real one after implantation, and its impact on the final dose distribution. AIM The aim of this study was to analyze the needle displacement as a function of prostate volume and provide precautions for the needle insertion procedure to improve the quality of the final dose distribution. MATERIALS AND METHODS In the Centre of Oncology MSC Memorial Institute Branch Gliwice patients with early stage localized prostate carcinoma (T1 T2) were treated with combined external beam therapy and HDR 192 Ir brachytherapy from Interstitial brachytherapy of the prostate is delivered as a 10 boost during a single treatment session. Treatment plans are based on transrectal ultrasound imaging and calculated with the Nucletron Swift treatment planning system, which was designed for real time conformal brachytherapy. On the basis of the transverse US images, taken with 1 mm scan thickness, volumes of interest such as PTV and organ at risk (urethra) are reconstructed. According to 3D VOI contours a virtual implant is prepared. Positions of the needles are established with coordinates of the template holes. Dwell positions in each catheter and dwell weights are determined in order to realize the optimal treatment plan. This virtual plan (Figure 1A) is evaluated with parameters of the dose volume histogram V 100 and D 10, which define respectively: the volume of prostate covered by 100% isodose surface, and the dose value that is delivered to 10% volume of the urethra. The implantation procedure is visualized on the live US images, which allow one to control live needle placement according to its virtual coordinates. Insertion of needles always begins with upper right horizontally and ends with lower left corner of the template. Even small differences between prescribed needle position and the real one result in changes in the virtual dose distribution. Therefore, on the basis of real implant geometry, dwell times are recalculated to obtain the final dose distribution ( live treatment plan) (Figure 1B). The quality of the live plan is also evaluated with dosimetric parameters V 100 and D 10. In this study thirty treatment plans ( virtual and live calculated for each patient) were randomly selected for analysis. For each treatment plan differences between virtual and real dose distribution caused by needle displacement were estimated by the new parameter VD (Volume Dose) (eq. 1) VD = V 100 virtual. D (1) 10 live V 100 live D 10 virtual

3 Rep Pract Oncol Radiother, 2008; 13(1): 9-14 Szlag M et al Prostate brachytherapy implant analysis A B Figure 1. Real-time conformal treatment planning for HDR brachytherapy of prostate cancer. (A) virtual treatment plan before implantation based on pre-planned catheter coordinates and anatomy-based optimization, (B) post-implant, final dose distribution calculated for real catheter positions and re-optimized for new implant geometry. Figure 2. The difference between virtual and real needle position r in mm measured on the selected axial image define needle displacement. V 100 virtual and V 100 live are fractions of the prostate volume that are encompassed by 100% isodose in the virtual and real treatment plan, while D 10 virtual and D 10 live are dose values that encompassed 10% volume of the urethra in the virtual and live treatment plan respectively. Values of the provided VD formula, which was created for the purpose of our analysis, characterize the final dose distribution with regard to the virtual plan: When: a. VD<1 real dose distribution is considered to be better than virtual one for prostate gland and/or urethra, b. VD>1 worsening in real dose distribution for prostate gland and/or urethra when compared to virtual plan, c. VD=1 virtual and real treatment plans are considered to be equivalent. Figure 3. Anatomic plane definition. After image acquisition base, reference and apex planes are defined. Apex refers to the plane that intersects the top of the prostate, while the base is the prostate plane closest to the bladder wall; reference is the middle and the largest cross-section of the gland. VD parameter was also designed to compare the differences between real and virtual treatment plans as a function of needle displacement and prostate volume. Needle displacement was determined by measuring the distance between virtual and real positions of respective needles (Figure 2) in three axial images (Figure 3) acquired at the base and the apex of the gland and in the reference image. Mean values of needle displacement were calculated for each plane and for three planes in total. RESULTS Needle number, needle displacement and prostate volume Number of implanted needles varied from 12 to 18 for single real treatment plan and the 11

4 Original Paper Rep Pract Oncol Radiother, 2008; 13(1): 9-14 Figure 4. Number of implanted needles depends on prostate volume Vp parameter. Number of needles and their initial coordinates are established with catheter placement option in treatment planning system on the basis of contoured patient anatomy. Virtual needle position and its number can be modified manually for better PTV coverage. Figure 6. Graph represents the scatter of VD values as a function of prostate volume Vp and needle displacement r. The values of r(vp) (solid line on the graph) boundary represent the maximum acceptable values of mean needle displacement r a and are relative to Vp. Analysis of VD parameter Influence of mean needle displacement and prostate volume on the final dose distribution was analyzed with VD parameter (Figure 6). In the group of treatment plans where VD 1 mean value of prostate volume was observed to be significantly higher than in the group of VD>1, while needle displacement was statistically equal in both groups (Table 1). Figure 5. Mean needle displacement for consecutive number of needle calculated for thirty treatment plans in three axial images. The biggest mean needle displacement was observed in the base plane of the prostate. median needle number was Mean prostate volume considered in this experiment was 29.75cc (±11.9cc). Needle number corresponded with prostate volume. Higher prostate volumes were implanted with higher needle number (Figure 4.) Mean value of the needle displacement r (Figure 2) calculated for thirty live treatment plans was 1.8mm (±0.6 mm) and did not depend on implantation sequence but varied depending on the level of transverse imaging: base, apex or reference. The needle displacement was the smallest for the apex and the biggest for the base plane (Figure 5). Changes in dosimetric parameters between virtual and live plan depend on needle displacement while prostate volume parameter seems to modulate the extent of dose distribution alteration. In order to obtain a real treatment plan at least equivalent to the virtual one, needle displacements for a given prostate volume are required to be equal to or smaller than acceptable, while the acceptable values for a mean needle displacement r are assumed to be determined by the expression: r acceptable (Vp) μ Vp 2 (2) Where: Vp prostate volume in cc by anatomy based contouring of PTV performed on the acquired US slice before implantation. r acceptable prescribed maximum acceptable values of needle displacement for a given Vp value. The expression led to the conclusion that for smaller glands even slight needle shift from 12

5 Rep Pract Oncol Radiother, 2008; 13(1): 9-14 Szlag M et al Prostate brachytherapy implant analysis Table 1. Differences between mean values of geometrical parameters in two groups of VD values. VD 1 n=16 VD>1 n=14 mean ± standard deviation p Mann-Whitney U test Mean needle displacement r [mm] 1.7± ± Prostate volume Vp [cc] 34.25± ± pre-planned positions may affect essential changes in virtual dose distribution that cannot be reduced by re-optimization. DISCUSSION Modern imaging techniques such as transrectal ultrasounds allow visualization of the procedure of implantation and on-line, real time evaluation of the accuracy of needle positioning according to its pre-planned position [7]. Not only for permanent seed implants, careful implant preparation and catheter geometry is an essential part of the treatment planning process in brachytherapy [8,9]. However, the potential tool of dwell time differentiation in HDR stepping source brachytherapy can minimize, to a certain degree, the superfluous underdosed and overdosed volumes resulting from needles displacement from their pre-planned positions. The research into the influence of an adequate needle placement regarding its virtual position as a function of prostate volume improves the possibility of predicting the quality of final dose distribution at the time of implantation. Charra-Brunaud et al. [10] suggested that a higher number of catheters leads to decreasing values of the V 150 parameter (percentage of prostate volume covered with 150% isodose) that allows large high-dose volumes inside the PTV to be avoided. 15 to 21 needles are an adequate number for preparing an optimal treatment plan. The results seems to be in agreement with our study where the number of implanted needles corresponds with prostate volume and was in the range of 12 to 18, which gave sufficient coverage of PTV by reference isodose in the virtual treatment planning phase. In our study the needle implantation sequence was taken into consideration. We assumed that after implantation of consecutive catheters prostate motion would reduce. According to Lagerburg et al. [2] the use of locking needles results in reduced prostate rotation in the coronal plane and thus improves the implantation accuracy. However, in our experiment no significant differences were observed between needle displacement for first and last implanted needles. Thus, during implantation, the prostate movements were not reduced by the increasing number of implanted needles. Implantation sequence or additional locking needles do not influence the reduction of prostate motion. The influence of mean needle displacement on dose distribution depends on prostate volume. Changes of the dose distribution calculated for bigger gland volume are less dependent on the needle shift than for smaller prostates (Figure 4). While mean needle displacement is higher than 2mm, final dose distribution for bigger prostate volumes (30cc) can be equivalent or even better than virtual when the renewed optimization procedure was applied to dose distribution calculated on the basis of real implant geometry. For prostate volumes lower than 25cc even small differences between virtual and real needle positions reduce the effectiveness of optimization methods. If the needle shift exceeds acceptable values for a given Vp parameter, determined by r a (Vp) function, needle relocation during implantation should be considered. Optimization algorithms are beneficial for conformal treatment planning in brachytherapy but their usefulness is restricted by insufficient implant geometry. CONCLUSIONS The calculations demonstrate that final dose distribution depends on needle displacement, which is the distance between virtual and real needle position, but prostate volume should be taken into consideration. The experiment determined maximal values of needle displacement for a given Vp parameter that still allow one to take advantage of optimizations algorithms to improve the final dose distribution. 13

6 Original Paper Rep Pract Oncol Radiother, 2008; 13(1): 9-14 However, acceptable values of needle displacement as a function of prostate volume require further research and more complex and multifactorial analysis. REFERENCES: 1. Kovacs G, Pötter R, Loch T et al: GEC/ESTRO- EAU recommendations on temporary brachytherapy using stepping sources for localized prostate cancer. Radiother Oncol, 2005; 74: Lagerburg V, Moerland MA, Lagendijk JJW, Battermann JJ: Measurement of prostate rotation during insertion of needles for brachytherapy. Radiother Oncol, 2005; 77: Spadinger I, Hilts M, Keyes M et al: Prostate brachytherapy postimplant dosimetry: A comparision of suture embedded and loose seed implants. Brachytherapy, 2006; 5: Sumida I, Shiomi H, Yoshioka Y et al: Optimization of dose distribution for HDR brachytherapy of the prostate using attraction-repulsion model. Int J Radiation Oncology Biol Phys, 2006; 64: Mate TP, Gottesman JE, Hatton J: High Dose Rate Afterloading 192 Ir Prostate Brachytherapy: Feasibility Report. Int J Radiation Oncology Biol Phys, 1998; 41: Martin T, Baltas D, Kurek R et al: 3-D Conformal HDR Brachytherapy as Monotherapy for Localized Prostate Cancer. A Pilot Study. Strahlentherapie und Onkologie, 2004; 180: Fijałkowski M, Białas B, Maciejewski B et al: Three Dimensional (3D) Real Time Conformal Brachytherapy A Novel Solution For Prostate Cancer Treatment. Part I. Rational And Method. Nowotwory Journal of Oncology, 2005; 55: Stone NN, Hong S, Lo Y-C et al: Comparison of intraoperative dosimetric implant representation with postimplant dosimetry in patients receiving prostate brachytherapy. Brachytherapy, 2003; 2: Potters L, Calugaru E, Jassal A, Presser J: Is there a role for postimplant dosimetry after real-time dynamic permanent prostate brachytherapy? Int J Radiation Oncology Biol Phys, 2006; 65: Charra-Brunaud C, Hsu I-C J, Weinberg V, Pouliot J: Analysis of interaction between number of implant catheters and dose volume histograms in prostate high dose rate brachytherapy using a computer model. Int J Radiation Oncology Biol Phys, 2003; 56: